Re: [gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Rossen Apostolov]

Hi,

On 9/13/11 4:27 PM, Mark Abraham wrote:

On 14/09/2011 12:20 AM, Marcin Zielinski wrote:

Ok,

Using -DGMX_ACCELERATION=Power6 brings a plethora of new errors 
during the compilation.


Firstly, including config.h inside the fortran .F kernel files for 
power6 is causing problems with
their parsing using xlf. adding -WF,-qfpp didn't help. Had to provide 
a modified xlf.cfg config file

for XLF (cppoptions = -P, instead of -C)
and pass it on with -F option and then add -qarch=pwr6 (all passed 
manually) for compilation of
each power6 kernel .F file. That fixed the problem with the 
compilation of Power6 kernels part.


Is there any more handy way of doing that? I've noticed that upon 
using Power6, cmake generates
Fortran_DEFINES and Fortran_FLAGS inside 
src/gmxlib/CMakeFiles/gmx.dir/flags.make file.
They somehow can not be modified from the command line. Same goes for 
the preprocessor flags.

(-DCMAKE_CPP_FLAGS do not work?).



You can also modify flags by editing CMakeCache.txt and rerunning cmake 
- if you don't pass options it should preserve what's been already found 
in the cache file.




Secondly, now I get something else:
Linking C shared library libmd.so
xlc: 1501-218 (W) file ../gmxlib/libgmx.so.6 contains an incorrect 
file suffix

../gmxlib/libgmx.so.6: In function `__bss_start':
(*ABS*+0x1025cb00): multiple definition of `_edata'
../gmxlib/libgmx.so.6: In function `__data_start':
(.data+0x8): multiple definition of `__dso_handle'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/crtbeginS.o:(.data.rel+0x0): 
first defined here

../gmxlib/libgmx.so.6: In function `_fini':
(.opd+0x30): multiple definition of `_fini'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(.opd+0x18): 
first defined here

../gmxlib/libgmx.so.6: In function `_init':
(.opd+0x18): multiple definition of `_init'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(.opd+0x0): 
first defined here

../gmxlib/libgmx.so.6: In function `_end':
(*ABS*+0x10268b20): multiple definition of `_end'
../gmxlib/libgmx.so.6: In function `__bss_start':
(*ABS*+0x1025cb00): multiple definition of `__bss_start'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/crtendS.o:(.dtors+0x0): 
multiple definition of `__DTOR_END__'

../gmxlib/libgmx.so.6:(.dtors+0x8): first defined here
/usr/bin/ld: error in ../gmxlib/libgmx.so.6(.eh_frame); no 
.eh_frame_hdr table will be created.

make[2]: *** [src/mdlib/libmd.so.6] Error 1
make[1]: *** [src/mdlib/CMakeFiles/md.dir/all] Error 2
make: *** [all] Error 2


Beats me, sorry.



me too :(

Rossen


Mark

--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Szilárd Páll]

Hi,

I have not followed the entire discussion so I might be completely
wrong, I might be fill in some gaps.

 Firstly, including config.h inside the fortran .F kernel files for power6 is
 causing problems with
 their parsing using xlf. adding -WF,-qfpp didn't help. Had to provide a
 modified xlf.cfg config file
 for XLF (cppoptions = -P, instead of -C)
 and pass it on with -F option and then add -qarch=pwr6 (all passed manually)
 for compilation of
 each power6 kernel .F file. That fixed the problem with the compilation of
 Power6 kernels part.

 Is there any more handy way of doing that? I've noticed that upon using
 Power6, cmake generates
 Fortran_DEFINES and Fortran_FLAGS inside
 src/gmxlib/CMakeFiles/gmx.dir/flags.make file.
 They somehow can not be modified from the command line. Same goes for the
 preprocessor flags.
 (-DCMAKE_CPP_FLAGS do not work?).

If you want to pass *global* flags to the the C++ compiler you can do
that through CMAKE_CXX_FLAGS variable (note: CXX not CPP!). However,
if you happen to need flags only for certain files or internal
libraries, it's slightly more tricky; it requires the file/target
property COMPILE_FLAGS to be modified fron inside the CMake scripts.

 Secondly, now I get something else:
 Linking C shared library libmd.so
 xlc: 1501-218 (W) file ../gmxlib/libgmx.so.6 contains an incorrect file
 suffix
 ../gmxlib/libgmx.so.6: In function `__bss_start':
 (*ABS*+0x1025cb00): multiple definition of `_edata'
 ../gmxlib/libgmx.so.6: In function `__data_start':
 (.data+0x8): multiple definition of `__dso_handle'
 /usr/lib64/gcc/powerpc64-suse-linux/4.3/crtbeginS.o:(.data.rel+0x0): first
 defined here
 ../gmxlib/libgmx.so.6: In function `_fini':
 (.opd+0x30): multiple definition of `_fini'
 /usr/lib64/gcc/powerpc64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(.opd+0x18):
 first defined here
 ../gmxlib/libgmx.so.6: In function `_init':
 (.opd+0x18): multiple definition of `_init'
 /usr/lib64/gcc/powerpc64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(.opd+0x0):
 first defined here
 ../gmxlib/libgmx.so.6: In function `_end':
 (*ABS*+0x10268b20): multiple definition of `_end'
 ../gmxlib/libgmx.so.6: In function `__bss_start':
 (*ABS*+0x1025cb00): multiple definition of `__bss_start'
 /usr/lib64/gcc/powerpc64-suse-linux/4.3/crtendS.o:(.dtors+0x0): multiple
 definition of `__DTOR_END__'
 ../gmxlib/libgmx.so.6:(.dtors+0x8): first defined here
 /usr/bin/ld: error in ../gmxlib/libgmx.so.6(.eh_frame); no .eh_frame_hdr
 table will be created.
 make[2]: *** [src/mdlib/libmd.so.6] Error 1
 make[1]: *** [src/mdlib/CMakeFiles/md.dir/all] Error 2
 make: *** [all] Error 2

That unfortunately looks like a nasty symbol multiple definition
issue. My guess is that the symbols xlc is complaining about are
defined in libgmx.so.6 and when the linker tries to do the final
linking stage it ends up choking due to conflicts with symbols is some
library where the respective symbols are defined (or in fact originate
from).

AFAIK this could happen when one provides a linker option
-lsomething to a compile-only command iso a [compile+] link command:
$ g++ -c foo.cpp -o foo.o -lbar
$ g++ foo.o -o foobar

Let me know if this makes sense, you could look at the verbose make
output (make VERBOSE=1) and maybe you notice something like this.

Cheers,
--
Szilárd

 cheers,
 --
 gmx-users mailing list    gmx-users@gromacs.org
 http://lists.gromacs.org/mailman/listinfo/gmx-users
 Please search the archive at
 http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
 Please don't post (un)subscribe requests to the list. Use the www interface
 or send it to gmx-users-requ...@gromacs.org.
 Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the
www interface or send it to gmx-users-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

[gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Marcin Zielinski]

Hi Mark,

I can confirm that the fix works.
Setting -DGMX_ACCELERATION=Power6 sets to these
/* Define to a macro mangling the given C identifier (in lower and upper
   case), which must not contain underscores, for linking with Fortran. */
#define F77_FUNC(name,NAME)  name

/* As F77_FUNC, but for C identifiers containing underscores. */
#define F77_FUNC_(name,NAME) name

which do not gives the mentioned error.

regards,

PS: I'll report back if I find something else.
--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

[gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Marcin Zielinski]

Ok,

Using -DGMX_ACCELERATION=Power6 brings a plethora of new errors during 
the compilation.


Firstly, including config.h inside the fortran .F kernel files for 
power6 is causing problems with
their parsing using xlf. adding -WF,-qfpp didn't help. Had to provide a 
modified xlf.cfg config file

for XLF (cppoptions = -P, instead of -C)
and pass it on with -F option and then add -qarch=pwr6 (all passed 
manually) for compilation of
each power6 kernel .F file. That fixed the problem with the compilation 
of Power6 kernels part.


Is there any more handy way of doing that? I've noticed that upon using 
Power6, cmake generates
Fortran_DEFINES and Fortran_FLAGS inside 
src/gmxlib/CMakeFiles/gmx.dir/flags.make file.
They somehow can not be modified from the command line. Same goes for 
the preprocessor flags.

(-DCMAKE_CPP_FLAGS do not work?).

Secondly, now I get something else:
Linking C shared library libmd.so
xlc: 1501-218 (W) file ../gmxlib/libgmx.so.6 contains an incorrect file 
suffix

../gmxlib/libgmx.so.6: In function `__bss_start':
(*ABS*+0x1025cb00): multiple definition of `_edata'
../gmxlib/libgmx.so.6: In function `__data_start':
(.data+0x8): multiple definition of `__dso_handle'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/crtbeginS.o:(.data.rel+0x0): 
first defined here

../gmxlib/libgmx.so.6: In function `_fini':
(.opd+0x30): multiple definition of `_fini'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(.opd+0x18): 
first defined here

../gmxlib/libgmx.so.6: In function `_init':
(.opd+0x18): multiple definition of `_init'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(.opd+0x0): 
first defined here

../gmxlib/libgmx.so.6: In function `_end':
(*ABS*+0x10268b20): multiple definition of `_end'
../gmxlib/libgmx.so.6: In function `__bss_start':
(*ABS*+0x1025cb00): multiple definition of `__bss_start'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/crtendS.o:(.dtors+0x0): multiple 
definition of `__DTOR_END__'

../gmxlib/libgmx.so.6:(.dtors+0x8): first defined here
/usr/bin/ld: error in ../gmxlib/libgmx.so.6(.eh_frame); no .eh_frame_hdr 
table will be created.

make[2]: *** [src/mdlib/libmd.so.6] Error 1
make[1]: *** [src/mdlib/CMakeFiles/md.dir/all] Error 2
make: *** [all] Error 2

cheers,
--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Mark Abraham]

On 14/09/2011 12:20 AM, Marcin Zielinski wrote:

Ok,

Using -DGMX_ACCELERATION=Power6 brings a plethora of new errors during 
the compilation.


Firstly, including config.h inside the fortran .F kernel files for 
power6 is causing problems with
their parsing using xlf. adding -WF,-qfpp didn't help. Had to provide 
a modified xlf.cfg config file

for XLF (cppoptions = -P, instead of -C)
and pass it on with -F option and then add -qarch=pwr6 (all passed 
manually) for compilation of
each power6 kernel .F file. That fixed the problem with the 
compilation of Power6 kernels part.


Is there any more handy way of doing that? I've noticed that upon 
using Power6, cmake generates
Fortran_DEFINES and Fortran_FLAGS inside 
src/gmxlib/CMakeFiles/gmx.dir/flags.make file.
They somehow can not be modified from the command line. Same goes for 
the preprocessor flags.

(-DCMAKE_CPP_FLAGS do not work?).

Secondly, now I get something else:
Linking C shared library libmd.so
xlc: 1501-218 (W) file ../gmxlib/libgmx.so.6 contains an incorrect 
file suffix

../gmxlib/libgmx.so.6: In function `__bss_start':
(*ABS*+0x1025cb00): multiple definition of `_edata'
../gmxlib/libgmx.so.6: In function `__data_start':
(.data+0x8): multiple definition of `__dso_handle'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/crtbeginS.o:(.data.rel+0x0): 
first defined here

../gmxlib/libgmx.so.6: In function `_fini':
(.opd+0x30): multiple definition of `_fini'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(.opd+0x18): 
first defined here

../gmxlib/libgmx.so.6: In function `_init':
(.opd+0x18): multiple definition of `_init'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(.opd+0x0): 
first defined here

../gmxlib/libgmx.so.6: In function `_end':
(*ABS*+0x10268b20): multiple definition of `_end'
../gmxlib/libgmx.so.6: In function `__bss_start':
(*ABS*+0x1025cb00): multiple definition of `__bss_start'
/usr/lib64/gcc/powerpc64-suse-linux/4.3/crtendS.o:(.dtors+0x0): 
multiple definition of `__DTOR_END__'

../gmxlib/libgmx.so.6:(.dtors+0x8): first defined here
/usr/bin/ld: error in ../gmxlib/libgmx.so.6(.eh_frame); no 
.eh_frame_hdr table will be created.

make[2]: *** [src/mdlib/libmd.so.6] Error 1
make[1]: *** [src/mdlib/CMakeFiles/md.dir/all] Error 2
make: *** [all] Error 2


Beats me, sorry.

Mark
--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Mark Abraham]

On 9/09/2011 1:46 AM, Marcin Zielinski wrote:

Hi there,

I've been having hard times the past days trying to build SP and DP 
binaries
of GROMACS 4.5.4 on our Power6 system, here at SARA, using IBM XL 
compilers,

LAPACK, BLAS, GSL and FFTW3.

At the end, I think I've succeeded and had to use two dirty hacks to 
avoid problems.

I'm waiting now for results confirmation from friend of mine.

First thing, small patch to cmake/FindLAPACK.cmake
-set(CMAKE_REQUIRED_LIBRARIES ${_flags} -Wl,--start-group
${${LIBRARIES}} ${_blas};-Wl,--end-group ${_threads})
+set(CMAKE_REQUIRED_LIBRARIES ${_flags} -Wl,--start-group
${${LIBRARIES}} ${_blas} -Wl,--end-group ${_threads})
(Note the removed semicolon. )


Thanks. Fixed for next version.



Create Your build directory, I had source dir and build dir in the 
same root dir.


Second thing, before issuing make (and after cmake) modify 
src/config.h inside Your build directory:
line 45: remove ## _ (should look like this: #define 
F77_FUNC(name,NAME) name)
line 48: remove ## _ (should look like this: #define 
F77_FUNC_(name,NAME)name)


If you set GMX_ACCELERATION=POWER6 this (and a few other things) should 
happen automatically.


Mark



After that, cmake command, for sp and tools only:
module load lapack fftw3/3.2.2 gsl/1.15 cmake

cmake \
  -DFFTW3F_INCLUDE_DIR=path to fftw3/include \
  -DFFTW3F_LIBRARIES=path to fftw3/lib/libfftw3f.a \
  -DGMX_EXTERNAL_BLAS=ON \
  -DBLAS_LIBRARIES=path to gsl/lib/libgslcblas.so \
  -DGMX_EXTERNAL_LAPACK=ON \
  -DLAPACK_LIBRARIES=path to lapack/lib/liblapack.a \
  -DGMX_GSL=ON \
  -DCMAKE_INSTALL_PREFIX=install dir/gromacs/4.5.4-sp \
  -DGMX_X11=OFF \
  -DCMAKE_CXX_COMPILER=mpCC \
  -DCMAKE_CXX_FLAGS=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 -qomp 
-qessl \
  -DCMAKE_CXX_FLAGS_RELEASE=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 
-qomp -qessl \

  -DCMAKE_C_COMPILER=mpcc \
  -DCMAKE_C_FLAGS=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 -qomp 
-qessl \
  -DCMAKE_C_FLAGS_RELEASE=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 
-qomp -qessl \

  -DCMAKE_F77_COMPILER=mpfort \
  -DCMAKE_F77_FLAGS=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 -qomp 
-qessl \
  -DCMAKE_F77_FLAGS_RELEASE=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 
-qomp -qessl \
  -DCMAKE_EXE_LINKER_FLAGS=-Lpath to fftw3/lib -L/usr/lib64 
-Lpath to gsl/lib -L/opt/ibmcmp/xlsmp/1.8/lib64 
-L/sara/sw/lapack/3.3.1-xlf12/lib -L/opt/ibmcmp/xlf/12.1/lib64 
-llapack -lgslcblas -lgsl -lesslsmp -lxlf90_r -lxlsmp -lxlfmath 
-lxlomp_ser  \
  -DCMAKE_SHARED_LINKER_FLAGS=-Lpath to fftw3/lib -L/usr/lib64 
-Lpath to gsl/lib -Lpath to xlf/xlsmp/1.8/lib64 -Lpath to 
lapack/lib -Lpath to xlf/xlf/12.1/lib64 -llapack -lgslcblas -lgsl 
-lesslsmp -lxlf90_r -lxlsmp -lxlfmath -lxlomp_ser \
  -DCMAKE_MODULE_LINKER_FLAGS=-Lpath to fftw3/lib -L/usr/lib64 
-Lpath to gsl/lib -Lpath to xlf/xlsmp/1.8/lib64 -Lpath to 
lapack/lib -Lpath to xlf/xlf/12.1/lib64 -llapack -lgslcblas -lgsl 
-lesslsmp -lxlf90_r -lxlsmp -lxlfmath -lxlomp_ser \

  -DGMX_MPI=OFF \
  ../gromacs-4.5.4

From here, there's a straight way for a parallel build for mdrun and 
same goes for double precision.

Hope this is going to help to anyone having similar problems as I did.

good luck,


--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

[gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Marcin Zielinski]

Hi Mark,

Yes, GMX_ACCELERATION should do the trick, however it requires You to 
turn off the THREADS.

Altho ppl tend to run the tools in serial anyway, I somehow sticked to that.

As for the FindLAPACK.cmake fix, I'm pretty sure I has been posted here 
some time ago. I googled it

anyhow.

Would be handy If I could put my hands on some benchmarking suit so that 
I could test this given
setup on my own and compare it to the build with POWER6 included (and 
without THREADS). Should I

see some boost in the speed in execution time?

best regards,
--
Marcin Zielinski
Supercomputing group, OSD
SARA
--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Mark Abraham]

On 9/09/2011 5:50 PM, Marcin Zielinski wrote:

Hi Mark,

Yes, GMX_ACCELERATION should do the trick, however it requires You to 
turn off the THREADS.
Altho ppl tend to run the tools in serial anyway, I somehow sticked to 
that.


Only mdrun takes advantage of parallelism in GROMACS up to now. So the 
non-threads requirement is moot for a build focused on the tools.




As for the FindLAPACK.cmake fix, I'm pretty sure I has been posted 
here some time ago. I googled it

anyhow.

Would be handy If I could put my hands on some benchmarking suit so 
that I could test this given
setup on my own and compare it to the build with POWER6 included (and 
without THREADS). Should I

see some boost in the speed in execution time?


Probably nothing much. There is no up-to-date benchmark suite.

Mark
--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

[gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Marcin Zielinski]

Ok, giving -DGMX_ACCELERATE=Power6 results with:

src/config.h
/* Define to a macro mangling the given C identifier (in lower and upper
   case), which must not contain underscores, for linking with Fortran. */
#define F77_FUNC(name,NAME)  name ##

/* As F77_FUNC, but for C identifiers containing underscores. */
#define F77_FUNC_(name,NAME) name ##

which, after issuing make, results with an error message at the very 
beginning:
/home/grmcs454/gromacsdir/test-build/src/config.h, line 45.39: 
1506-210 (S) The ## operator requires two operands.
/home/grmcs454/gromacsdir/test-build/src/config.h, line 48.39: 
1506-210 (S) The ## operator requires two operands.


I checked cmake/CMakeLists.txt and '##' are put explicitly therefor they 
will be present with _
afterwards or without it, thus giving this error every time You put 
Power6 or Fortran option on.

As far as I understand.


Only mdrun takes advantage of parallelism in GROMACS up to now. So the
non-threads requirement is moot for a build focused on the tools.

--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Re: [gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Mark Abraham]

On 9/09/2011 10:01 PM, Marcin Zielinski wrote:

Ok, giving -DGMX_ACCELERATE=Power6 results with:

src/config.h
/* Define to a macro mangling the given C identifier (in lower and upper
   case), which must not contain underscores, for linking with 
Fortran. */

#define F77_FUNC(name,NAME)  name ##

/* As F77_FUNC, but for C identifiers containing underscores. */
#define F77_FUNC_(name,NAME) name ##

which, after issuing make, results with an error message at the very 
beginning:
/home/grmcs454/gromacsdir/test-build/src/config.h, line 45.39: 
1506-210 (S) The ## operator requires two operands.
/home/grmcs454/gromacsdir/test-build/src/config.h, line 48.39: 
1506-210 (S) The ## operator requires two operands.


I checked cmake/CMakeLists.txt and '##' are put explicitly therefor 
they will be present with _
afterwards or without it, thus giving this error every time You put 
Power6 or Fortran option on.

As far as I understand.


They make sense if $suffix is non-empty, but it can be empty, so it 
looks like someone didn't code this right. If the patch below works, let 
me know and I'll fix the source version.


Mark

diff --git a/CMakeLists.txt b/CMakeLists.txt
index f26734d..a59a1de 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -636,13 +636,24 @@ if(GMX_FORTRAN OR GMX_POWER6)
 if(prefix)
   set(prefix ${prefix} ##)
 endif(prefix)
+if(suffix)
+  set(suffix ## ${suffix})
+  if(extrasuffix)
+   set(extrasuffix ${suffix}${extrasuffix})
+  endif(extrasuffix)
+else(suffix)
+  if(extrasuffix)
+   # Don't know if this is needed, but it can't hurt
+   set(extrasuffix ## ${extrasuffix})
+  endif(extrasuffix)
+endif(suffix)

 if(isupper)
-set(F77_FUNCDEF   ${prefix} NAME ## ${suffix})
-set(F77_FUNCDEF_  ${prefix} NAME ## ${suffix}${extrasuffix})
+set(F77_FUNCDEF   ${prefix} NAME ${suffix})
+set(F77_FUNCDEF_  ${prefix} NAME ${extrasuffix})
 else(isupper)
-set(F77_FUNCDEF   ${prefix} name ## ${suffix})
-set(F77_FUNCDEF_  ${prefix} name ## ${suffix}${extrasuffix})
+set(F77_FUNCDEF   ${prefix} name ${suffix})
+set(F77_FUNCDEF_  ${prefix} name ${extrasuffix})
 endif(isupper)
 else(GMX_FORTRAN OR GMX_POWER6)
 set(F77_FUNCDEF   name ## _)

--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

[gmx-users] building GROMACS 4.5.4 on Power6 with CMAKE

[Marcin Zielinski]

Hi there,

I've been having hard times the past days trying to build SP and DP binaries
of GROMACS 4.5.4 on our Power6 system, here at SARA, using IBM XL compilers,
LAPACK, BLAS, GSL and FFTW3.

At the end, I think I've succeeded and had to use two dirty hacks to 
avoid problems.

I'm waiting now for results confirmation from friend of mine.

First thing, small patch to cmake/FindLAPACK.cmake
-set(CMAKE_REQUIRED_LIBRARIES ${_flags} -Wl,--start-group
${${LIBRARIES}} ${_blas};-Wl,--end-group ${_threads})
+set(CMAKE_REQUIRED_LIBRARIES ${_flags} -Wl,--start-group
${${LIBRARIES}} ${_blas} -Wl,--end-group ${_threads})
(Note the removed semicolon. )

Create Your build directory, I had source dir and build dir in the same 
root dir.


Second thing, before issuing make (and after cmake) modify src/config.h 
inside Your build directory:
line 45: remove ## _ (should look like this: #define F77_FUNC(name,NAME) 
name)
line 48: remove ## _ (should look like this: #define 
F77_FUNC_(name,NAME)name)


After that, cmake command, for sp and tools only:
module load lapack fftw3/3.2.2 gsl/1.15 cmake

cmake \
  -DFFTW3F_INCLUDE_DIR=path to fftw3/include \
  -DFFTW3F_LIBRARIES=path to fftw3/lib/libfftw3f.a \
  -DGMX_EXTERNAL_BLAS=ON \
  -DBLAS_LIBRARIES=path to gsl/lib/libgslcblas.so \
  -DGMX_EXTERNAL_LAPACK=ON \
  -DLAPACK_LIBRARIES=path to lapack/lib/liblapack.a \
  -DGMX_GSL=ON \
  -DCMAKE_INSTALL_PREFIX=install dir/gromacs/4.5.4-sp \
  -DGMX_X11=OFF \
  -DCMAKE_CXX_COMPILER=mpCC \
  -DCMAKE_CXX_FLAGS=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 -qomp 
-qessl \
  -DCMAKE_CXX_FLAGS_RELEASE=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 
-qomp -qessl \

  -DCMAKE_C_COMPILER=mpcc \
  -DCMAKE_C_FLAGS=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 -qomp -qessl \
  -DCMAKE_C_FLAGS_RELEASE=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 
-qomp -qessl \

  -DCMAKE_F77_COMPILER=mpfort \
  -DCMAKE_F77_FLAGS=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 -qomp 
-qessl \
  -DCMAKE_F77_FLAGS_RELEASE=-V -q64 -O3 -qarch=pwr6 -qtune=pwr6 
-qomp -qessl \
  -DCMAKE_EXE_LINKER_FLAGS=-Lpath to fftw3/lib -L/usr/lib64 
-Lpath to gsl/lib -L/opt/ibmcmp/xlsmp/1.8/lib64 
-L/sara/sw/lapack/3.3.1-xlf12/lib -L/opt/ibmcmp/xlf/12.1/lib64 -llapack 
-lgslcblas -lgsl -lesslsmp -lxlf90_r -lxlsmp -lxlfmath -lxlomp_ser  \
  -DCMAKE_SHARED_LINKER_FLAGS=-Lpath to fftw3/lib -L/usr/lib64 
-Lpath to gsl/lib -Lpath to xlf/xlsmp/1.8/lib64 -Lpath to 
lapack/lib -Lpath to xlf/xlf/12.1/lib64 -llapack -lgslcblas -lgsl 
-lesslsmp -lxlf90_r -lxlsmp -lxlfmath -lxlomp_ser \
  -DCMAKE_MODULE_LINKER_FLAGS=-Lpath to fftw3/lib -L/usr/lib64 
-Lpath to gsl/lib -Lpath to xlf/xlsmp/1.8/lib64 -Lpath to 
lapack/lib -Lpath to xlf/xlf/12.1/lib64 -llapack -lgslcblas -lgsl 
-lesslsmp -lxlf90_r -lxlsmp -lxlfmath -lxlomp_ser \

  -DGMX_MPI=OFF \
  ../gromacs-4.5.4

From here, there's a straight way for a parallel build for mdrun and 
same goes for double precision.

Hope this is going to help to anyone having similar problems as I did.

good luck,
--
Marcin Zielinski
Supercomputing group, OSD
SARA
Science Park 140   mob +31 (0)65 123 5602
1098 XG Amsterdam  tel +31 (0)20 888 4074
www.sara.nlfax +31 (0)20 668 3167
marcin.zielin...@sara.nl
--
gmx-users mailing listgmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to gmx-users-requ...@gromacs.org.

Can't post? Read http://www.gromacs.org/Support/Mailing_Lists